Review of progress, QNDE

Elastic wave scattering at grain boundaries in polycrystalline media can be quantified to determine microstructural properties. The amplitude drop observed for coherent wave propagation (attenuation) as well as diffuse-field scattering events have been extensively studied. In all cases, the scattering shows a clear dependence on grain size, grain shape, and microstructural texture. Models used to quantify scattering experiments are often developed assuming dependence on a single spatial length scale, usually, mean grain diameter. However, several microscopy studies suggest that most metals have a log normal distribution of grain sizes. In this study, grain size distribution is discussed within the context of previous attenuation models valid for arbitrary crystallite symmetries. Results are presented for titanium using a range of distribution means and widths assuming equiaxed grains and no preferred crystallographic orientation. The longitudinal and shear attenuations are shown to vary with respect to the frequency dependence for varying distribution widths even when the volumetric mean grain size is held constant. Furthermore, the results suggest that grain size estimates based on attenuation can have large errors if the distribution is neglected. This work is anticipated to play an important role in microstructural characterization research associated with ultrasonic scattering.

Ultrasonic attenuation of polycrystalline materials with a distribution of grain sizes.

One limiting factor in the detection of internal defects in metals is so‐called “grain noise” which arises from the scattering of sound by microstructural boundaries For pulse/echo inspections, independent‐scatterer models have been developed and applied successfully to predict backscattered grain noise characteristics in many situations of practical importance Here we consider the extension of the independent‐scatterer noise models to pitch‐catch geometries In this first of several planned papers, we focus on the evaluation of the scattering coefficient η which is a property of the microstructure alone and quantifies its capacity for generating grain noise A general scheme for evaluating η is first laid out Then, for simple single‐phase cubic and hexagonal metals, formulas are developed which express η in terms of the single‐crystal elastic constants and the mean grain diameter for various combinations of the incident and scattered wave types

Computation of Grain‐Noise Scattering Coefficients for Ultrasonic Pitch/Catch Inspections of Metals

In past work we reported on measurements of ultrasonic velocity, attenuation and backscattering in nickel‐alloy materials used in the fabrication of rotating jet‐engine components. Attenuation and backscattering were shown to be correlated to the average grain diameter, which varied with position in the billet specimens studied. The ultrasonic measurements and associated metallographic studies found the local microstructures to be approximately equiaxed and free of texture in these cubic‐phase metals. In this paper we explore a method for deducing the single‐crystal elastic constants of a metal using the combined ultrasonic and metallographic data for a polycrystalline specimen. We specifically consider the case seen in the jet‐engine alloys: polycrystalline cubic microstructures having equiaxed, randomly oriented grains. We demonstrate how the three independent elastic constants {C11, C12, C44} can be deduced from the density, the mean grain diameter, the ultrasonic attenuation at one or more frequencies, and the longitudinal and shear wave speeds. The method makes use of the attenuation theory of Stanke and Kino, and the Hill averaging procedure for estimating the sonic velocity through a polycrystalline material. Elastic constant inputs to the velocity and attenuation models are adjusted to optimize the agreement with experiment. The method is demonstrated using several specimens of Inconel 718 and Waspaloy, and further tested using four specimens of pure Nickel.

Estimation of Single‐Crystal Elastic Constants from Ultrasonic Measurements on Polycrystalline Specimens

For pulse/echo inspections of metals, models which predict backscattered noise characteristics often make a “single‐scattering” assumption, i.e., multiple‐scattering events in which sound is scattered from one grain to another before returning to the transducer are ignored. Models based on the single‐scattering assumption have proven to be very useful in simulating inspections of engine‐alloy billets and forgings. However, this assumption may not be accurate if grain scattering is too “strong” (e.g., if the mean grain diameter and/or the inspection frequency is too large). In this work, backscattered grain noise measurements and analyses were undertaken to search for evidence of significant multiple scattering in pulse/echo inspections of jet‐engine Nickel alloys. At or above about 7 MHz frequency and 50 micron grain diameter, problems were seen with single‐scattering noise models that are likely due to the neglect of multiple scattering by the models. The modeling errors were less severe for focused‐probe measurements in the focal zone than for planar probe inspections. Single‐scattering noise models are likely adequate for simulating current billet inspections which are carried out using 5‐MHz focused transducers. However, multiple scattering effects should be taken into account in some fashion when simulating higher‐frequency inspections of Nickel‐alloy billets having large mean grain diameters (> 40 microns).

Looking for Multiple Scattering Effects in Backscattered Ultrasonic Grain Noise from Jet‐Engine Nickel Alloys

Ultrasonic velocity, attenuation, and back‐scattered grain noise have been measured in rectangular coupons cut from representative 10″‐diameter billets of IN718 and Waspaloy. Ultrasonic attenuation and backscattered noise were found to vary significantly with position within a billet, principally with radial depth. However, at a given measurement site there was little dependence of ultrasonic properties on inspection direction, suggesting an approximately equiaxed, untextured microstructure. Subsequent metallographic examinations revealed equiaxed grain structures in which the average grain diameter varied with position, tending to be largest at sites with large attenuations and large grain noise levels. The manner in which attenuation or backscattered‐noise capacity (FOM) grows with increasing average grain diameter is similar to that expected for Pure‐Ni microstructures. However, the rise rates are somewhat smaller for the jet‐engine alloys, likely due to differences between the single‐crystal elastic constants of the alloys and those of pure Ni. This paper reviews the methods used for ultrasonic measurements and metallographic analyses, and summarizes the interrelationships between attenuation, backscattered noise capacity and average grain diameter.

F. J. Margetan

Papers

Correlation between Local Ultrasonic Properties and Grain Size within Jet-Engine Nickel Alloy Billets